[BUG]Edit Config on local EKS cluster

[rolfbartels]

Rancher Server Setup

• Rancher version: 2.92
• Installation option (Docker install/Helm Chart): Helm
  • If Helm Chart, Kubernetes Cluster and version (RKE1, RKE2, k3s, EKS, etc): EKS
• Proxy/Cert Details: Rancher Cert from Cert Manager

Information about the Cluster

• Kubernetes version: 1.30
• Cluster Type (Local/Downstream): local

• Hosted = EKS

User Information

• What is the role of the user logged in? Admin/Cluster Owner

Describe the bug
On a new install or old install on Rancher 2.9.2 I want to add labels to the local cluster.
When I edit the config I am prompted that atleast 1 node group must be configured.
I insert the missing field (Cluster Name, Cloud Creds (Same Creds use to eksctl create), add a node group and try and save.

I am presented on and error
{"data":""}

To Reproduce
Cluster Management
Select the ellipsis for local cluster
Edit Config
Enter in required details
Save

Result
{"data":""}
Expected Result
Cluster Config saves

Screenshots

Additional context

[rak-phillip]

@rolfbartels thanks for raising this issue and helping to make Rancher better! Our team might follow-up for to collect more information as we triage the issue.

@gaktive transferring to the Dashboard repo for triage. This could be related to SURE-9216

[rak-phillip]

I can repro this issue in v2.9.3. The form behaves exactly as described when editing the local cluster for Rancher hosted on EKS

[gaktive]

@mantis-toboggan-md this feels like something we can tackle in 2.11.0 and 2.10.1 but can you take a quick look to see how easy a fix could be?

[mantis-toboggan-md]

@gaktive I have looked into this and spoken about it a little with Phil. Investigation is ongoing; we don't yet know if this is a frontend, backend, or combined bug. Given that, I'd say it's not a good candidate for 2.10.0.

[rak-phillip]

Following up with some information based on OOB discussions. We've identified that the behavior described in the issue is related to eksStatus.upstreamSpec evaluating to null and eksConfig is undefined. This has been reported to be expected behavior:

on local cluster where Rancher is running, and even if it was deployed on EKS eksStatus should be:

eksStatus:
  generatedNodeRole: ''
  managedLaunchTemplateID: ''
  managedLaunchTemplateVersions: null
  privateRequiresTunnel: null
  securityGroups: null
  subnets: null
  upstreamSpec: null
  virtualNetwork: ''

it is because local cluster is not provisioned or managed by eks-operator

Based on the information provided, I think that we will need to define the edge-cases for working with local clusters and handle them accordingly so that the form behaves as expected for this use case.

[gaktive]

/backport v2.10.1

[mantis-toboggan-md]

/backport v2.9.5